Authors' response

We would wish to thank Dr Mackin for pointing out our error in stating that the Pope et al (1991) study used valproate semisodium and not sodium valproate. He also draws attention to a possible misinterpretation in our section on tolerability. We did not wish to imply that Epilim Chrono, a combination formulation of sodium valproate and valproic acid, was enteric coated. However we believe this does not alter our underlying opinion that US data derived from non-enteric coated forms of valproate should not be extra-polated in relation to the general use of valproate in the UK, unless interpreted cautiously. Epilim Chrono by nature of its modified release formulation does not release a bolus of valproate in the stomach and based on the assumption that gastric side effects are concentration related, cannot be regarded as the same as non-enteric coated valproate with respect to gastric side effects. Although Epilim Chrono may be the most commonly used preparation in the UK, figures from the Prescription Pricing Bureau in England show that it only accounts for around 30% of valproate items dispensed in tablet form.

At the time we submitted our paper, 2003 data mentioned by Dr Mackin were not available to us. Dr Mackin mentions the NICE (2003) Technology Appraisal, number 66. We were surprised by the government-directed remit for this particular appraisal, set out in section 3.1; that only medicines with a licence for the treatment of bipolar disorder were to be considered. This is in contrast to other areas of medicine where NICE have clearly felt ‘off licence’ use was within their remit. One effect of this was to exclude drugs with an established ‘off licence’ use and create a spurious belief among many psychiatrists that valproate semisodium had to be used in preference to other forms of valproate because it was the only form approved of by NICE. Local Drug and Therapeutics Committees were left to struggle with this issue.